The effects of chloride-based deicing chemicals on degradation of portland cement mortars with alkali reactive aggregate

The assessment of the influence of deicing chemicals on alkali reactivity of aggregates is still an area of active research, particularly with respect to the combined effects of long-term exposure to deicers, freezing and thawing cycles, and wetting and drying cycles. In order to develop a more comprehensive view of these topics, the present study was conducted in three phases. The focus of the first phase was on studying the effects of the chloride-based deicers on the alkali silica reaction (ASR) using accelerated test methods (i.e. the test was performed at high temperature (80 °C) as per modified ASTM C1260). The second phase focused on the effect of deicers on the ASR under prolonged exposure to freezing and thawing (F/T) and wetting and drying (W/D) cycles. The aim of the third phase was on establishing the e effects of chloride-based deicing chemicals on the aggregates alone (i.e. without the presence of hydrated paste) using both, the reactive and non-reactive aggregates and exposing them to deicers at 4 °C, 23 °C and 80 °C. The silica sand (nearly pure quartz from Ottawa, IL) was used as (nominally) non-reactive aggregate and Jobe sand (mixed quartz/chert/ feldspar material from near El Paso, TX) was used as a reactive aggregate in this study. The soak solutions used in the experiments were DI water, 1 N NaOH, and three chloride-based deicers (NaCl, MgCl2 and CaCl2) with nearly-eutectic concentrations. The results of Phase 1 of the study showed that the mortar bars with reactive aggregate (Jobe sand) stored in NaCl solution developed higher expansion than those stored in MgCl2 and CaCl2 solutions. The examination of the microstructure using the scanning electron microscope (SEM) revealed that exposure of mortar with reactive aggregate to NaCl initiated and accelerated the ASR at early ages and that there was no evidence that MgCl2 and CaCl2 contributed to ASR. However, after relatively short (about 150 days) period of exposure to MgCl2 solution at 80 °C, mortar bars started to disintegrate, irrespective of whether they contained reactive or non-reactive aggregates. This was attributed to the chemical attach of the deicer on the cementitious matrix and formation of large amounts of gypsum within the microstructure. The results of Phase 2 showed that the specimens exposed to CaCl 2 expand most in the F/T test while the specimens exposed to MgCl 2 expanded most in the W/D test. The microstructural analysis of the specimens revealed that the expansion of the mortars was mainly caused by formation of the oxychloride phases and that it was related to microcracking induced by the F/T cycles. The specimens exposed to MgCl2 had the highest reduction of compressive strength, followed by the specimens exposed to CaCl2. On the other hand, the compressive strength of the specimens exposed to NaCl in the W/D test was even slightly higher than the strength of the reference samples exposed to deionized (DI) water. Even the prolonged exposure (up to 245 F/T and W/D cycles) did not produce significant evidence of ASR during the tests period. The results of Phase 3 demonstrated that, when exposed at 80 °C, the reactive Jobe sand was capable to raise the pH value of the DI water, thus increasing the likelihood of initiating the ASR. Similarly, the level of Na+ ions in the 1N solution of NaOH in contact with Jobe sand at 80 °C continued to decrease over time, confirming that the material was undergoing ASR. On the other hand, in the absence of cement paste the exposure to deicers did not produce any significant evidence of ASR. The overall conclusion from the study was that under certain exposure conditions the deicers with alkali ions (i.e. NaCl) can increase the potential for ASR in the systems with reactive aggregates whereas the effect of the other deicers is more related to the chemical attack on the components of the hydrated paste